Re: [Mingw-users] [PATCH] Support MSYS2

[Peter Rosin]

On 2015-07-24 23:32, Greg Jung wrote:
> On Fri, Jul 24, 2015 at 2:03 PM, Peter Rosin <address@hidden 
> <mailto:address@hidden>> wrote:
>> 
>>     [ mingw-users: this is in reply to a suggested patch (1) ]
>> 
>>     On 2015-07-23 19:30, address@hidden <mailto:address@hidden> wrote:
>>     > Hello,
>>     > when using the MSYS2 shell (http://msys2.github.io/), uname -o returns 
>> Msys instead of mingw32. The latest versions of config.guess and config.sub 
>> already support that, but Libtool doesn't. Here's a patch to add that 
>> support.
>> 
>>     This does not feel correct as a quick glance. Are you sure you
>>     understand what you are doing?
>> 
>>     In normal operation, I would have assumed that msys2 returns
>>     mingw-something and that any msys mode was a special case for
>>     building binaries with the cygwin/posix emulation layer, but
>>     you are adding msys to a bunch of places which are mingw-only.
>>     So, it all feels fishy.
>> 
>>     Also, how does this affect msys1?
>> 
>>     Cheers,
>>     Peter
>> 
>>     (1) http://lists.gnu.org/archive/html/libtool/2015-07/msg00003.html
> In all three of my msys2 XXX_shell.bat shells, uname -o answers Msys.
> in msys-1, uname -o answers Msys
> 
> The query that gets you to mingw32 is "uname -s"
> 
> In MSYS2:
> address@hidden MSYS ~
> $ uname -s
> MSYS_NT-6.1
> 
> address@hidden MINGW32 ~
> $ uname -s
> MINGW32_NT-6.1
> 
> address@hidden MINGW64 ~
> $ uname -s
> MINGW64_NT-6.1
> 
> In Mingw/MSYS from mingw.org <http://mingw.org>:
> address@hidden ~
> $ uname -s
> MINGW32_NT-6.1
> 
> In Cygwin it looks like so:
> 
> address@hidden ~
> $ uname -o
> Cygwin
> 
> address@hidden ~
> $ uname -s
> CYGWIN_NT-6.1

Oh, "uname -o". I didn't notice that. But I also do not see any reference
to "uname -o" in the latest config.sub/config.guess. So, someone is
confused. Is it me?

Cheers,
Peter